Jump to ContentJump to Main Navigation
The Pragmatics of Politeness$

Geoffrey Leech

Print publication date: 2014

Print ISBN-13: 9780195341386

Published to Oxford Scholarship Online: August 2014

DOI: 10.1093/acprof:oso/9780195341386.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

A Case Study

A Case Study


(p.115) 5 A Case Study
The Pragmatics of Politeness

Geoffrey Leech

Oxford University Press

Abstract and Keywords

This chapter focuses on a particular type of speech event manifesting politeness: the apology. First, the nature of speech-event categories such as apology is investigated. There are rarely all-or-nothing distinctions in politeness behavior; thus apologies are manifested in scales of gradience and are found in nonprototypical as well as prototypical instances. Apologies are potentially more than a simple speech act: they can involve a main or head act such as I’m sorry, and also subsidiary acts such as an admission of guilt, or an explanation of why the fault occurred. Politeness is also characteristic of respons to apologies, manifesting the neg-politeness of the Maxim of Obligation by the hearer, and thereby contrasting with the pos-polite Maxim of speaker’s Obligation that is typically found in apologies. Finally, the chapter gives attention to public apologies.

Keywords:   apology, I’m sorry, prototypical, responses, Maxim of Obligation, neg-politeness

Research on politeness has tended to concentrate on particular types of speech events or social encounters that are “politeness-sensitive,” and there is an advantage in focusing on these in accounting for a wide range of politeness behavior, much of it highly conventionalized. The next three chapters will focus on these speech events. We concentrate in this chapter on apologies, in Chapter 6 on requests, and in Chapter 7 on other speech events, such as offers, compliments, and thanks, as well as on responses to these speech events.

I refer to these as speech events rather than as speech acts, because the latter term has typically been used in the study of single utterances, a particular limitation of Searle’s speech act theory (1969, 1975a). However, when we study such phenomena as requests and apologies in context, we often find that they are more complex than this. For example, Blum-Kulka and her associates in the CCSARP project1 find that apologies are part of a complex of individual acts or moves. These are listed below as (a)–(e) in order of their frequency in and centrality to apologies. Although apologies are frequently single-move and even single-word events like Sorry! (in which case we can still call them speech acts), the term more generally can apply to a whole complex of IFIDs (illocutionary force indicating devices) and accompanying “satellite” moves.

This chapter and the next focus on two speech events—apologies and requests—that offer particularly salient demonstrations of politeness in English-speaking societies and therefore deserve close attention in themselves. An additional reason for (p.116) focusing on them is to supply exemplary “case studies” illustrating a more detailed plan of analysis than will be possible for other speech events.

5.1 Apologies: Speech Events Seen as Prototype Categories

The most important of Blum-Kulka et al.’s potential components of an apology is known as a head act, which is more essential than other components, and contains the nub or nucleus of the speech event:

  1. (a) head act: the apology itself (IFID), e.g.: (I’m) (so) sorry...

Another component of speech act realizations, in CCSARP, is the occurrence of modifiers (see Blum-Kulka et al. 1989, Trosborg 1995), which qualify the meaning of the speech event especially by intensifying or softening its force in some degree. An adverb like very or terribly can obviously act as a modifier intensifying the apology, as in I’m very sorry; I’m terribly sorry. This is an example of pos-politeness (see 1.2.2). But in speech events of neg-politeness, such as requests, the modifier is more likely to be a hedge or downgrader: Could I just have a little slice?

These intensifiers, downgraders, and the like are known as internal modifiers because they are syntactically incorporated into the main components of the apology.2 There are also external modifiers: CCSARP applies this term, for example, to exclamatory elements such as Oh dear! (preceding apologies) or (in requests) alerters such as Excuse me or vocatives such as Barry, love. These usually stand apart from the head act, either preceding or following it, being syntactically describable as isolates (see 3.3). However, they can still contribute to the politeness of the speech event, by adding attitudinal or emotive meaning.

In addition, an apology may be accompanied by satellite speech events:

  1. (b) a confession, or admission of responsibility for the fault, such as (I’m sorry, ) I lied. [LCSAE 094501]

  2. (c) an explanation of why or how the fault occurred: (Well I’m sorry it’s been such a mess.) It’s just, this, this whole magazine this year, I mean we’ve had to rely on so many other folks, you know [LCSAE 125202]. (Goffman 1971: 109–115 refers to such explanations or excuses as “accounts”.)

  3. (d) an offer of repair (or making amends): making sure the fault is to be corrected or a remedy applied; e.g., (after spilling something) (Rachel, I’m sorry I’m leaving this here.) I’ll tidy it up. [BNC KNR]

  4. (e) promise of forbearance (making amends in the longer term by undertaking to do better on future occasions): (Right, right, so I’m very (p.117) sorry.) I won’t do it again next year. [BNC HE2] (spoken by a lecturer apologizing to students for lack of book availability)

These five moves are subdivided by the CCSARP team into the head act (a), and the supporting moves (b)–(e). Also, CCSARP calls (a) and (b) general components of an apology, and (c), (d), and (e) specific components more associated with particular types of apology. How far (c)–(e) are likely to occur depends very much on the situation. For example, if you have just spilled coffee on someone’s new jacket, the promise not to do it again (e) is unlikely to be mollifying. However, an offer of repair (d), such as undertaking to have the garment cleaned immediately, would be more acceptable, and more in keeping with the goal of propitiating the victim. As for (c), if S has done something to inconvenience H, such as missing an appointment, then it will be felt S “owes” H an explanation. It will be felt insufficient if S says I’m sorry I didn’t meet you at three o’clock without saying why the mishap occurred.

As this CCSARP analysis already hints, speech events such as apologies and requests do not constitute clear-cut categories. They are more like prototype categories (Rosch 1977, 1978; Ungerer and Schmid 2006: 7–63). A proper on-the-record apology can be recognized as the prototype, at the top end of a scale of typicality, as distinct from a more peripheral example of an apology such as (1)3:

  1. (1) Oh ooh, I kept you waiting.

  2. (2) I’m sorry I kept you waiting... [BNC JXW]

Let us recognize that (2) is obviously a more forthright apology than (1), which lacks the IFID I’m sorry, and so might almost be considered a nonapology, merely mentioning that a fault has occurred. But even I’m sorry does not signal an apology unambiguously; it has to be placed in a suitable context, linguistic or extralinguistic. For example, in another example (3), component (b) above, the “self-blame” element of an apology, is not present at all, or is only questionably present.

  1. (3) I’m sorry that our time is up, but many thanks to my three guests today...[BNC KRL]

The speaker is the anchor in a local BBC radio program, expressing regret that the program has to come to an end. There is surely no reason for the speaker to accept blame for the BBC timetabling that resulted in the end of the program. (It could be, admittedly, that the speaker is considering herself to be the mouthpiece of the BBC, which has committed a fault in bringing the program to a premature end. But this seems implausible.) Rather, S is expressing regret that she has to do something that may cause some disappointment to the listener, without imputing self-blame.

The prototype view of the apology is championed by Deutschmann (2003: 46), who links it to a semantic frame consisting of four components.

A Case StudyApologies

Figure 5.1 The four components included in a prototypical apology.

Note: Roughly based on Figure 2.2 in Deutschmann (2003: 46).

(p.118) According to Deutschmann, the components can be expanded on as follows:

  1. 1. The offender, who takes responsibility for the offense but did not necessarily cause it

  2. 2. The offended, who is perceived to have suffered as a result of the offense

  3. 3. The offense—real, potential, or perceived as such by the offender or the offended

  4. 4. The remedy—recognition of the offense, acceptance of responsibility, and a display of regret

Deutschmann then goes on to mention three kinds of apology that are nonprototypical, “which fall partly outside this prototypical view of the speech act” (ibid.):

  1. (a) “Formulaic apologies” where the “offense is minimal,” and where apologizing is more or less a matter of routine, e.g., saying sorry for “social gaffes such as coughing, slips of the tongue.”

  2. (b) “Formulaic apologies” with added functions, where the offense “is minimal and has other functions in addition to repair work,” e.g., cueing a request, or calling for attention: “Sorry?” “Excuse mecould you pass that microphone?”

  3. (c) “Face attack apologies,” as Deutschmann calls them (cf. also Culpeper 2011a: 174–178). These typically preface a speech event that is likely to be seen as impolite, such as a directive or a refusal (violating the Tact Maxim) or a complaint or criticism (violating the Approbation Maxim), as in:

  1. (4) I’m sorry but I just think that’s outrageous. [BNC KPU]

  2. (5) Your mum is mad, I’m sorry but she is. [BNC KPH]

Arguably, as Deutschmann points out, these apologies have rather little, if any, remedial effect, as S, having apologized in advance, goes on to commit the face-threatening act.

(p.119) I’m sorry is often used as a preface to an FTA in this way, and S’s justification for her own behavior may be signaled by a following but. However, but may also indicate that an FTA follows:

  1. (6) Well I’m sorry but my Daily Mirror divi—[was] delivered this morning without the television supplement...[BNC KB8]

The I’m sorry could be elaborated as follows: “I’m sorry that I’m going to say something unpleasant to you...” and its effect at most is to slightly mitigate the impoliteness of the FTA. The but is also significant, and can be explicated as follows: “I’m sorry, but despite that, I’m going to say the unwelcome thing that I have to say.” The oddity in (6) is that a “real” offense has been committed by H (or H’s co-workers), who failed to ensure that S got her TV supplement; but S feels it necessary to apologize for making the complaint.

Another reason for nonpoliteness in apologies is that one of the supporting moves to an apology, (c), the explanation of the offense, often serves a self-excusing purpose, making the offense seem smaller. In this sense, explanations may help to save S’s face:

  1. (7) Yes, look I’m sorry we’ve got to raise rents by nine and a half percent. Unfortunately the government is assuming that’s all we are doing so they’re cutting housing subsidy accordingly and we would have been left with no choice...[BNC JT8]

Nevertheless it can be argued that the explanation (here beginning with Unfortunately...) contributes to the mitigation of the offense, as without it the offense would seem far more blameworthy, and the apology, even though it may be judged insufficient, can do something to make the bad news more acceptable. Notice that many of these self-excuses, as in (7), make use of deontic modality (must, have to, ’ ve got to, etc.), making the offense seem beyond S’s control.

This discussion has shown that the most common means of expressing an apology, (I’m) sorry, is actually more of a variable signal, not always signifying an apology, and not always conducive to politeness. Searle’s IFIDs are not always as watertight, as might be supposed. This underlines the need to consider categories like “apology” as prototypes rather than rule-governed categories as in classic speech act definitions.

5.2 A Digression: Apologies and Other Speech Events

So, rather than thinking of speech events such as apologies as forming a clear-cut illocutionary type, it is helpful to think of them as covering a particular “illocutionary territory” with internal variations as well as contrastive relations with other speech events. To help us to think in these terms, I am presenting Table 5.1, based on the same grid as was seen in Table 4.1 in the last chapter, in which apologies are located (p.120) in relation to the maxims of the GSP, but adding two other parameters: indicating whether the speech event exemplifies pos-politeness or neg-politeness, and whether it is S(peaker)-oriented or O(ther)-oriented.

Table 5.1 The territorial relations of various politeness-sensitive speech events.

Maxims of the GSP

Typical speech event types

Pos- or neg-



S- or


M1. Generosity

M2. Tact

Offering, inviting, promising

Requesting, ordering, entreating





M3. Approbation

M4. Modesty

Complimenting, praising

Responding to compliments, etc.





M5a. Obligation (S to O)

M6a. Obligation (O to S)

M5b. Obligation (S to O)

M6b. Obligation (O to S)


Responding to thanks


Responding to apologies









M7. Agreement

M8. Opinion reticence


Disagreeing, advising





M9. Sympathy

M10. Feeling reticence

Congratulating, comforting, well-wishing

Responding to congratulations, etc





The right side of this table (highlighted in bold) indicates the two factors just mentioned: first, the distinction between pos- and neg-politeness discussed in 1.2.2, and second, the distinction between S-oriented and O-oriented speech events.

Briefly, the best clue to pos-politeness is to test whether intensifying modifiers can be added or further intensified to increase the degree of (pragmalinguistic) politeness. As exemplified below, in (i) paying a compliment, (ii) thanking someone for a favor,4 (iii) expressing agreement, or (iv) expressing sympathy, such intensification is the most obvious way to make one’s speech act more polite:

  1. (i) Thanks a lot. | I’m extremely grateful. | Thank you very much indeed.

  2. (ii) That suits you perfectly. | Thanks for a wonderful meal.

  3. (iii) I totally agree with you. | Absolutely. | I couldn’t have put it better myself.

  4. (iv) I was so sorry to hear about your...| Many many congratulations. | Have a great time.

In contrast, neg-politeness becomes (pragmalinguistically) more polite through the use of hedges or downgraders. This applies, for example, to requests (see 7.3.3) such as Wouldyou mind just being quiet for a moment? to negating responses to thanks (It was nothing), to apologies (No problem, It doesn’t matter a bit), to disagreements, etc.

(p.121) What about apologies? By this criterion, they are examples of pos-politeness. To increase the (pragmalinguistic) politeness of an apology we intensify, making the apology appear more genuine, and the regret more profound:

  1. (8) I’m really very, very sorry.

  2. (9) We most sincerely apologize...

But if the question is asked, “What kind of speech event is an apology: pos-polite or neg-polite?” then instinctively we might feel it must be neg-polite—since its main function is one of repair: to repay the debt, to redeem S’s loss of face. However, it is the intended effect of a speech act on H that is crucial: an apology is meant to be face-enhancing to H rather than face-threatening.5 Harking back to Table 4.1 (p. 91), an apology is a transaction that gives value to O. That is, by apologizing, S not only acknowledges a fault but pays a debt to the face of O.

The right-hand column in Table 5.1 marks each speech event as either S-oriented or O-oriented. This again is a fairly clear criterion to operate with. First, we recall (from 1.2.1) that each of the speech events listed embodies or refers to some kind of transaction, that is, some social action carried out between S and O. The question here is whether the action is carried out by (or whether the event involves) the speaker or the other person—here H. And again, speech events tend to group themselves into pairs of opposites, for example, offer versus request, thanks versus apology. Thus an offer is S-oriented because it envisages a favor to be performed by the speaker; whereas a request is O-oriented as it envisages a favor to be performed by O, the other person. A thank-you is O-oriented, presupposing a (usually previous) action by the other person; whereas an apology is S-oriented, presupposing a (usually previous) action by the speaker. Notice, however, that S- and O-orientation can be manipulated in the way a speech event is presented. This can be done, for example, by disguising an O-oriented speech event as if it were S-oriented: thus a request can be presented in the form Can I borrow this printer? in contrast to Can you lend me this printer? A further step in disguise is to avoid personal references to “I” and “you” entirely, using a strategy of impersonality: Is it OK to borrow this lamp? A significant observation is that in the BNC Can I borrow...? is seven times more frequent than Can you lend...? This may be a sign that by avoiding direct reference to O, S disguises the demand imposed on O and hence makes the request appear less face-threatening. A somewhat similar switch from S-orientation to O-orientation can take place with apologies:

  1. (10) Oh I’m sorry I forgot about that other class. [LCSAE141701] (S-oriented)

  2. (11) I’m sorry you had to wait [LCSAE145502] (apparently O-oriented)

In (11) the speaker, apparently a teller, is apologizing for a delay at the bank, but focusing on the inconvenience for H rather than on whose fault it is.

(p.122) 5.3 Prototypical and Less Prototypical Apologies

Searle’s classic definitions of speech acts in terms of felicity conditions, however, cannot be entirely discarded. They neatly summarize what is central, in semantic terms, to an apology. Table 5.2 presents a blueprint for an apology, using Searle’s four kinds of conditions, and modeling them on his definition (1969: 67) of the related speech act of thanking.

Table 5.2 A definition of an apology modeled on Searle (1969)

Propositional content condition

Past act A by S

Preparatory condition

H is harmed by A and S believes H is harmed by A

Sincerity condition

S acknowledges responsibility to A and feels regret for A

Essential condition

Counts as an expression of contrition for A

Regarding this as a prototype definition, though, it is possible for each of the conditions to be waived under particular circumstances. Let’s consider each condition in turn.

The condition that is most likely to be waived is the sincerity condition, since, as we know, people sometimes speak insincerely: violations of the Maxim of Quality do occur. In fact, an apology is a speech event that is very easy to perform insincerely. It is easy for me to say I apologize without anyone knowing whether I feel any twinge of regret (whereas, for example, if I make an insincere promise, it will be found out when the promised action is not performed).

Moving on to the propositional content condition: this can be partially waived by locating the offending act A in the future or present, rather than in the past. I’m sorry to interrupt, spoken in the very act of interrupting, is an instance of an apology where A takes place at the same time as the apology—indeed, where A and the apology are the same event. Many everyday apologies for mild offenses take the form of an Excuse me... or Pardon me... where the directness of the imperative may be an overt sign that no serious offense or face threat is being admitted. Anticipatory (future-pointing) apologies are usually for minor offenses, where S does not expect any objection to hinder the performance of the act. Excuse me can be used as an “alerter”6 to catch the attention of a stranger whom S wants to engage in conversation, or to give warning that S is going to intrude on H’s physical or mental space in some way. It is noticeable that the force of these imperatives is so (p.123) low-key that please is either not added at all (in the case of Pardon me) or is added only very rarely (in the case of Excuse me).7 Notice that if a more polite request formula is used, as in Would you excuse me (please)? the meaning of excuse me implies a more serious offense:

  1. (12) I don’t wish to be rude, would you excuse me I have a bus to get to [BNC KRX]

This kind of utterance is typically used when S is about to break off from some ongoing social activity, such as a meeting, a conversation, or a meal, so the meaning in (12) is roughly ‘Would you allow me to leave?’ Another partial waiver of the propositional content condition occurs where the offending act A is conditional, and presented in an if-clause, which, in contrast to a that-clause, is nonfactive:

  1. (13) I’m sorry if it wasn’t clear. [LCSAE125301]

  2. (14) I’m sorry if my behavior wasn’t up to scratch. [FLOB P06]

As S does not definitely admit the fault, this could be considered a qualified or less-than-complete apology. It can be somewhat offensive to put an apology in this conditional form, especially if O thinks S is at fault; compare (14) with the factivity of I’m sorry (that) my behaviour wasn’t up to scratch, where S definitely confesses the fault.

As we have already seen, the preparatory condition is also waived, where S does not acknowledge culpability but merely expresses regret. There is no doubt that S in (15) both expresses regret and implies an offense:

  1. (15) I’m sorry to bother you [LCSAE 163001]

But in (16) S expresses regret with no likely imputation of guilt:

  1. (16) I’m sorry it’s raining. [BNC KRM]

Yet notice that (16) can still serve as an apology of a kind: S can still reasonably apologize for bad weather where H’s exposure to bad weather is due to some act by S, such as making arrangements for an outdoor event, or inviting H to visit S’s home on a rainy day. There is a kind of “guilt by association,” where S does not commit any offense but nevertheless accepts some responsibility for what happened. One kind of apology that is interesting in this respect is a mutual apology, where both parties apologize for the same mishap. It quite often happens, for example, in British society, where the people jostle each another in the street, that a quick “Sorry”—“Sorry” is exchanged. Although this is in the category that Deutschmann calls “formulaic apologies...where the offence is minimal,” a significant thing is that (p.124) both interlocutors seem to assume guilt, however small. This is a striking example of the reciprocal asymmetry of politeness (2.1b): in line with the PP (Maxim of Obligation), each person claims the fault to be theirs, rather than the other person’s.

The essential condition (“A counts as an expression of contrition”) is so closely bound up with the preparatory condition in this case8 that the examples I have given in (15) and (16) can also be used to illustrate waiver of the essential condition; (16) in particular hardly counts as an expr ession of contrition.

5.4 Apologies: The Pragmalinguistic Facet

In analyzing the nature of apology as a speech event, we have to pay attention to both the pragmalinguistic and sociopragmatic facets of analysis (1.2.3 and 1.2.4). But first let’s concentrate on a level that interacts with both these: the level of speech event strategies, which, as the term suggests, concerns what techniques (semantically speaking) are adopted for conveying an apology. A great deal of apologizing is done by means of syntactically condensed formulaic devices; the formulae (I’m) sorry, pardon, excuse me often occur alone. Here a high degree of pragmaticalization has taken place, which means that these formulae have acquired “bleached”9 (or weakened) meanings, and reduction through ellipsis is very common—for example Sorry is much more common than I’m sorry in the BNC spoken material. Similarly, Pardon is much more common than (I) beg your pardon or Pardon me in both the LCSAE and the BNC. One of the surest signs of pragmaticalization is that an expression loses its grammatical status as (say) a verb or a noun and becomes a mere pragmatic particle, as has already been noted (1.2.3) regarding please. This seems to have happened with Pardon, which in present-day English could be regarded as a noun (a reduced version of I beg your pardon) or as a verb (a reduced version of Pardon me). Forms like this seem to be canonical cases of what Watts (2003) calls “politic behavior,” which is neither polite nor impolite but simply follows convention. Yet the apologetic meaning of such formulae is easy to recover from their form, and they still retain some of the pragmatic value of politeness markers, as we see by comparing their interrogative use as requests for repetition (Sorry? Excuse me? Pardon?) with such functional equivalents as What? Huh? or Eh? which are often felt to be impolite.10

(p.125) Apologetic formulae used as requests for repetition, such as Sorry? are another example of “guilt by association.” That is, by apologizing for failure to hear what H said, one is taking on oneself the guilt that could equally well be attributed to H, for failing to speak clearly or loud enough.

Bearing in mind that a large majority of apologies are routine and formulaic, it can still be said that English uses three main (semantic) strategies of apology, listed here in order of frequency of use:

  1. (A) EXPRESSION OF S’S REGRET: e.g., (I’m) sorry, I regret..., I’m afraid...

  2. (B) ASKING H’S PARDON (OR FORGIVENESS): e.g., Excuse me, Pardon (me)

  3. (C) USING A PERFORMATIVE UTTERANCE: e.g., I/We apologize, I beg your pardon

(Less common reduced formulae, such as the noun phrase (My) apologies, sometimes intensified as in (My) deepest apologies, are rare enough to be ignored here.) We now briefly examine the various lexicogrammatical forms of apology, as listed in (A)–(C) above.

5.4.1 Sorry

We begin with strategy (A). Sorry (whether or not preceded by I’m) is by far the most common expression for apology in English, and of the various structural possibilities of sorry the use of it as an isolate (with no syntactic connection with other elements) is the most common. Two frequent usages are (a) declarative (typically followed by a full stop in the BNC transcription) and (b) interrogative (typically followed by a question mark). The intonation for (a) is typically a low rise or a fall rise, and for (b) typically a high rise. These variants are respectively illustrated in (17) and (18).

  1. (17) ‘You don’t have a lighter by any chance, do you?’

    ‘I’m afraid I don’t smoke. Sorry.’ [BNC JJS]

  2. (18) ‘Do you wanna drink while we’re here?’


    ‘Would you care for a drink?’ [BNC K71]

Example (19) shows another common variant, similar to (17) except that the order is different: here sorry is followed, rather than preceded, by a closely linked utterance that gives the reason for the apology:

  1. (19)Sorry, I didn’t ask if you wanted any more tea.’ [BNC KPV]

(p.126) Examples (20) to (22) illustrate the most common structurally linked patterns in which Sorry occurs.

  1. (20) Sorry about your orange juice. 〈laugh〉 Got a little thirsty...[LCSAE 147402]

  2. (21) I’m sorry I’m so disorganized. [LCSAE 141902]

  3. (22) Sorry to keep bothering you. [LCSAE 138403]

Sorry about... as in (20) identifies an offense in the reduced form of a noun phrase—often leaving much to what can be inferred from the context (the inexplicit phrase Sorry about that being the most common realization of this pattern). In (21) sorry is followed by a that-clause specifying the offense (in this case with ellipsis of that). The to-infinitive construction following Sorry in (22) is a typical way of apologizing for S’s current behavior (examples such as Sorry to be a nuisance and Sorry to interrupt are typical). But with the perfect infinitive, it can also be a way of apologizing for a past offense, as in this example from a parish meeting:

  1. (23) Thank you chairman. Sorry to have laboured the point. [BNC H49]

In contrast to sorry, I/we regret introduces a formal apology, which is more likely to occur in a written text rather than in speech:

  1. (24) PI Associates, sole manufacturers of PolyPads, wish it to be known that Bridlesuite of Lanarkshire have never been supplied with or ever stocked genuine PolyPads. We regret any inconvenience this may have caused. [BNC ASH]

Like I’m sorry, however, I/we regret does not necessarily signal an apology. The self-blame element is clearly absent from an example like (25):

  1. (25) If you produce a proposal, the first thing a lot of British engineers will do is tell you what’s wrong with it. That, I regret, is a British characteristic. [BNC A6L]

I’m sorry and I’m afraid can both be followed by a content clause, which however has a rather different function in the two constructions. I’m afraid is more like an apologetic preface to a statement presented as new information; an example is found in (17) above—I’m afraid I don’t smoke—where it is clear that “I don’t smoke” is treated as a fact previously unknown to H. In contrast, consider the clause following I’m sorry in (21), I’m sorry I’m so disorganized. Here the exclamatory statement I’m so disorganized expresses a presupposition, conveying information already available to both S and H. Notice that replacing I’m sorry by I’m afraid here would produce a barely acceptable utterance: ?I’m afraid I’m so disorganized.

5.4.2 Excuse me / Pardon (me)

Moving on to strategy (B), we first observe that these imperative formulae Excuse me and Pardon (me) are largely restricted to routine apologies. The imperative verb excuse, (p.127) compared with the verb forgive or pardon, seems to mean that H is enjoined not so much to forgive as to make allowances for S’s less-than-acceptable conduct. Excuse me is therefore largely concerned with mild offenses. It can apply to physical misdemeanors like sneezing, coughing, belching, and yawning, or infringements of conversational proprieties such as interrupting another speaker’s turn, or interrupting a piece of conversation in order to do something else, as in answering the phone in (26):

  1. (26) Excuse me a wee second, Charlie [answering the phone] [BNC G5E]

  2. (27) Excuse me while I find our papers...[BNC FM2]

  3. (28) Excuse me for being a bit thick...[BNC HUV]

In (28) S is apologizing (probably ironically) for her inability to understand what O has said. In AmE (American English) Excuse me and Pardon me can have additional functions such as signaling a speech error or disfluency (29) or—in the interrogative—asking for repetition (30):

  1. (29) ...it was due the day after the spring break, or excuse me, Christmas break. [LCSAE 127402]

  2. (30) ‘That’s wild.’ ‘Excuse me?’ ‘That’s wild’. [LCSAE 094501]

Pardon (me) has similar functions to Excuse me, as in asking for repetition:

  1. (31) Pardon, I didn’t hear the question? [LCSAE 153602]

5.4.3 I Apologize, I Beg Your Pardon

Strategy (C), the performative strategy, is infrequent compared with the other two. I apologize occurs mainly as a formal on-record apology, which (perhaps because there is no mention of regret or any other emotive involvement) can carry an air of “apology for formality’s sake only,” or (as in the case of (32)) of humor:

  1. (32) And here I apologize for the rather poor sound quality, but he refused to shut his parrot and dogs in another room during the interview. [BNC KRH]

  2. (33) erm I know there are many familiar faces around so I apologize to those of you that may know some of this information already. [BNC J3W]

A common manifestation of this performative apology is the conditional apology (comparable to I’m sorry if... ; see (13) in 5.3), which is a stock formula on public road signs and the like:

  1. (34) We apologize for any inconvenience this may (have) cause(d).

Because of the nonfactivity of any, as well as the modal may, such notices cost little and admit to no offense.

(p.128) (I) Beg your pardon is chiefly a British form (in the United States it is mainly used for mock indignation). Like Pardon, it is often used in asking for repetition, as in this example from an interview:

  1. (35) A: Have you ever had a job?

    B: Beg your pardon?

    A: Have you ever had a job? [BNC J8F]

The performative I/we apologize can be softened by turning it into a hedged performative (see 6.4.1(e)), using a modal expression:

(36) I should apologize for canned soup, but I spent so long talking to Stanley this morning. [BNC KC0]

This is no longer an on-record apology, and it could in fact be considered a nonapology, as the modal auxiliary converts the apology into an aspiration (what I should do) rather than a fact. Other hedged performatives have the declarative form with must or have to (e.g., I must apologize for..., expressing an obligation) or the interrogative form with can or may (e.g., Can I apologize...? asking permission). Each conveys the implication that S is willing and ready to apologize, but they do not actually enact an apology. Another version of a hedged apology is illustrated in (36), a quotation from a newspaper:

  1. (37) ‘Once again I can only apologise on behalf of my clients for events that occurred.’ [BNC K54]

5.5 Apologies: The Sociopragmatic Facet

The sociopragmatic facet relates the various kinds and degrees of apologies to the social circumstances in which they occur. We can look at these both from the viewpoint of various factors within English-speaking societies and from the external point of view—examining the incidence of apologies in English-speaking societies as compared with other language communities. This section can only touch on the subject.

Holmes’s study of apologies among New Zealand speakers of English (1990, 1995) found a great difference between females and males. Females apologized to others much more, and also received more apologies from others. Holmes (1995: 1) comes to the conclusion that women are more polite than men, in apologies as well as other speech events (however, this analysis has been challenged by others such as Mills 2003: 213–214). Holmes’s analysis of the four dyadic types—women apologizing to women, men to men, women to men, and men to women—shows a pleasing systematicity: woman-to-woman apologies are the most frequent, men-to-men apologies the least frequent, with women-to-men and men-to-women apologies showing an intermediate frequency. The tendency for women to give more tokens of politeness than men is illustrated by these studies.

(p.129) However, Deutschmann, pointing out that Holmes’s studies use only small samples, analyzes apologies in a large corpus—the BNC—and reaches the unexpected conclusion that males apologize more than females! He also finds, less surprisingly, that working-class speakers apologize less, not being in the “politeness culture” of the middle classes.

Another surprising finding of Deutschmann’s study, obviously connected with his finding that men apologize more than women, is that more powerful people apologize more to less powerful people than vice versa. This is at odds with Brown and Levinson’s well-known claim that politeness increases in proportion to three factors: the power and social distance of H relative to S, and the weight of the imposition (or what is transacted).

Turning to sociopragmatic differences between nationalities or regional groups, we find it useful to consider the five sociopragmatic parameters mentioned in 4.4.5: vertical distance, horizontal distance, and cost-benefit (corresponding to B&L’s P, D, and R factors), supplemented by the two extra factors of “strength of obligations/rights” and “self-territory vs. other territory.” Tanaka (1991) recalls—shortly after her arrival in Australia—going into a store to return a faulty desk lamp just purchased, and being shocked by the storekeeper’s response I see. Do you want to exchange it? rather than the profuse, embarrassed apology that would have greeted such a complaint in Japan. The sense of obligation and deference of the service provider to the customer in Japan is famously much greater than in English-speaking countries, which explains the very different norms of polite behavior in Japan and in Australia in this kind of situation. Another anecdote recounted by Tanaka (ibid.) is of a Japanese visitor to Australia whose car was damaged by an accident caused by a young Australian driver. Visiting the driver’s home to seek compensation, he was appalled that the driver’s parents did not find it necessary to apologize. Here the variable factor of “self-territory vs. other-territory” seems to be at issue: for the Japanese, the child is part of the extended “self-territory” of the parents, so the debt or fault of the child naturally belongs also to the parent. These are rather extreme cases of different apology norms; but the contrast between English-speaking and Japanese society is more typically a matter of degree; for example, Tanaka (ibid.) found that vertical and horizontal distance had more influence on Japanese students’ apologizing behavior than on Australian (English-speaking) students.

These differences between Japanese and Australian apology behavior can more generally be attributed to the distinction—or rather the cline—between individualist and collectivist societies. The former—typically “Anglo” or Western societies—prize the independence of the individual, whereas the latter—of which Japan is an example—give importance to the mutual dependence of members of a hierarchically organized social group. This can explain, in part, how the Japanese service provider shows great deference to the customer’s needs, as the service provider or storekeeper is seen less as an individual than as a representative of the trading firm.

(p.130) Perhaps the variation in the differences of apologetic behavior among language communities can be overestimated. This was certainly the conclusion of Olshtain (1989) in her study of apologies within the CCSARP research group. She found that her subjects representing speakers of four language communities behaved very similarly in apologizing (or not apologizing) for a set of offenses. However, the four languages concerned (English, French, German, and Hebrew) are all associated with Europe or a European cultural heritage. Also, as Olshtain pointed out, the various situations tested were limited, because of the need for comparability, to a student’s life in a Western society, where we might expect differences between one national group and another to be minimized.

5.6 Responses to Apologies

While polite speech events tend to come in contrasting pairs paradigmatically—as in the case of requests and offers or thanks and apologies (see Table 5.1)—they also tend to come in pairs syntagmatically; that is, one speech event typically evokes, or is followed by, another kind of speech event as response. The two events constitute an adjacency pair (Schegloff 2007). Here there is an enactment of the asymmetry of politeness (see 2.1b). A serious apology, in expressing S’s guilt and regret, by implication invites H to reply by either accepting or rejecting the apology. As with other politeness-sensitive speech acts, there is a preference for the polite option: an absolving speech event by which S is excused, or pardoned, for the offense. This can be done by denying the reality of the offense—No offense taken, No problem, It was nothing—or by implying that the fault was trivial enough to be ignored—It doesn’t matter; Think nothing of it; Don’t mention it; Forget it; etc. Another common forgiving response, That’s/It’s all right/OK, seems to accept the apology as sufficient recompense for the fault.

  1. (38) A [lecturer]: Hello.

    B: Hello I’m sorry I’m late.

    A: That’s all right. [BNC FUE]

Of course, other, non-exonerating responses are possible but dispreferred (see 2.1e):

  1. (39) A [pupil]: I’m sorry I’m late

    B [teacher]: You don’t look as if you mean it, sit down please because my lesson started ten minutes ago. [BNC F7R]

It is no doubt significant that the forgiving reply in (38) is addressed by the lecturer of the introductory lecture of a preretirement course to a new arrival—who is likely to be of relatively advanced years and not previously known to the speaker (factors tending to increase politeness), whereas the unforgiving response in (39) is spoken by a teacher to a child, where the vertical distance in favor of the speaker (p.131) is great. Both of these replies, preferred and dispreferred, are however marked in comparison with the normal situation, where the apology is routine and highly conventionalized, and no response to it is felt necessary:

  1. (40) Welcome to the Archaeological Resource Centre. Sorry about the delay, 〈pause〉 but the school in front of you were nearly half an hour late. Thank you very much for turning up on time. Now so welcome to the ARC.

Notice that the speaker pauses after the apology in this speech, but no one takes the opportunity to register any kind of verbal response (although smiles and nods might have hinted to him that he was forgiven).

5.7 Public Apologies

This is a suitable point to mention a genre of “political apologies” in public life that has become prominent in the media in recent decades. “The Art of the Public Grovel,” as Wise Bauer (2009) calls it in her book of that name, has been extensively analyzed and commented on in academic publications; for pragmatic treatments relevant to politeness, see Abadi (1990), Brooks (1999), Liebersohn et al. (2004) and Harris et al. (2006). The sequence of events is usually this: a public scandal breaks in the media, whereby some public figure is accused of a previously concealed offense, which may be against an individual or a large community. As an example, I quote here short snippets from a transcript of Tiger Woods’s public apology (on February 19, 2010) for marital unfaithfulness, addressed to his wife, his family, his friends, and his fans, which has been available on his website (http://web.tigerwoods.com/news/article/201002198096934/news/).

The head act, saying sorry, is clear: For all that I have done, I am so sorry.... There is also an admission of responsibility, I recognize I have brought this on myself..., and also an offer, or promise, of repair: I owe it to those closest to me to become a better man. That’s where the focus will be. In fact, the full transcript contains several examples of such moves. The one supporting move that is not easy to identify is an explanation of why the offense occurred: the “account.” The nearest thing to such an explanation was his confession that his fame and fortune had led him astray: I stopped living by the core values I was taught to believe in....I thought I could get away with whatever I wanted to.... This is the nearest one is likely to get to a full, thoroughgoing apology in public. Even so, some commentators were critical of it and dubbed it a “semi-apology.” Perhaps this simply illustrates the prototype nature of speech events like apologies (as discussed in 5.1): the criteria for judging what is and what is not an apology, a promise, and so on, are rarely clear-cut.

Four curious features of the public apology, as a genre, are that first it takes place in the full glare of the media (although it may be written or spoken). Second, both S and O can be a large number of people. Third, the actual deliverer of the (p.132) apology may not be directly involved in the offense but may be in some sense representative of the body that was responsible for it (an oil company apologizing for pollution; Catholic bishops apologizing for abuse of children by priests, or for failing to deal severely enough with such offenders). In extreme cases, the apologizer may not have been alive at the time of the offense (e.g., the Japanese prime minister apologizing for use of women of other countries as sex slaves in World War II). And fourth, the apology is usually in response to public (including media) complaints and is designed to assuage public outrage, especially bearing in mind that more severe consequences, such as resignations, lawsuits, or criminal proceedings might be incurred. When the apologizer is speaking on behalf of a public body, however, a full-blooded apology (including the use of the A word itself) is often avoided, as it implies liability in a legal sense and could lead to untold financial loss. In expressing “regret,” on the other hand, S does not himself or herself necessarily admit to the offense. In this respect the apology may be an incomplete one. It is worth remembering that considerations of politeness are not the main motivations governing the public use of apologies. Usually there is a political, social, or financial motivation too. The admission of guilt that accompanies a full apology can have negative legal, political, or financial consequences, as well as leading, in certain circumstances, to severe loss of face.

What public apologies bring to the fore is that apologizing, more strongly than other politeness-sensitive speech events, is face-threatening for the speaker. Following Ruhi (2006), Spencer-Oatey et al. (2008: 111) point out a limitation of my politeness framework, as well as that of B&L, in that it has a “bias towards ‘concern for other’,” whereas “self-presentation is another important interactional concern that needs to be incorporated into any explanatory account of the management of relations/rapport.” In other words, speech events like compliments, requests, and apologies have to take into account regard for S—what Spencer-Oatey et al. call self-presentation—as well as regard for O. I admit that this cannot be handled in my politeness model, which explicitly deals with communicative altruism, and is therefore inevitably focused on “concern for other.” But I have also recognized that this politeness model is not meant to be an account of the totality of what Spencer-Oatey called “rapport management.” I have in addition noted that politeness cannot only be face-saving and face-enhancing for O, but as a secondary effect can be so for S. However, what we find with apologies, especially public apologies, is that they can involve (severe) loss of face for the speaker, and they may be avoided or toned down for that reason. The positive reason for apologizing in order to restore a balance of good relations has to be weighed against the negative reason for avoiding apology in order to avoid face loss or humiliation.

5.8 Conclusion

In conclusion, taking the apology as an example of a speech event, it may be useful to list the various factors we have noticed as defining the nature of what we may call “apology territory,” and the topography of that territory, in terms of what variations occur.

(p.133) Apology Territory

Politeness characteristics (as listed in Table 5.1): Maxim of Obligation (of S to O); pos-politeness; S-oriented

An apology speech event consists of: a head act (IFID) (a)

and possible supporting moves: an expression of responsibility (b)

an explanation of why the fault occurred (c)

an offer of repair (d)

a promise of forbearance (e)

Sincerity conditions: S accepts responsibility for the fault, and feels regret for it

Related speech event(s): Thanking

Head act formulae and expressions: (I’m) sorry; (I’m) sorry about..., / to.../ (that) ...; I beg your pardon; I apologize; pardon me; excuse me; forgive me; pardon; (my) apologies

Internal modifications: intensifying: very/so/really etc. sorry; I apologize profusely; please excuse me; etc.

External modifications: discourse markers, names, etc. Oh; Oh dear; John (etc.)

Polite adjacency pair: apology followed by acceptance/exoneration.

Variation of strength: mild apologies (excuse me, etc.); deep apologies (... really sorry, etc.)

Other variations: retrospective, simultaneous, anticipatory apologies; conditional vs. actual apologies; personal vs. public (e.g., political) apologies

Apologies typically refer to past events; naturally, one might think, only an offense that has already been committed can be excused. However, we have already seen that Excuse me can be used to apologize for an offense S is committing at the time of speech (e.g., interrupting), or even for a future offense S plans to commit:

  1. (42) I’m sorry I can’t make it to the meeting tonight.

The failure to attend the meeting is in the future, but S already has made plans to commit the “offense.” The excuse, however, is that S will not do this deliberately; the use of can’t in (42) implies that her absence is unavoidable. Another nonprototypical type of apology is the conditional apology represented by the semi-formulae I’m sorry if... and We apologize for any..., which could be said not to be “real” apologies, as S does not own up to committing an offense. The implication is that S does not know whether an offense was committed or not.

After Chapter 6 on requests, Chapter 7 will deal in outline with a range of speech events, including thanks (7.3), which are closely related to apologies; and commiserations or condolences (7.5.3), which overlap to some extent with apologies, since I’m sorry can denote S’s regretful feeling for reasons of sympathy rather than guilt. Both emotions, of course, can be simultaneously involved.


(1) This chapter owes much to previous schemes of speech act analysis, such as have been provided by the Cross-Cultural Speech Act Realization Project (CCSARP; see Blum-Kulka et al. 1989 and 2.2.6) and other researchers the field, such as (on requests and apologies) Trosborg (1995) and Schauer (2009). Specifically, on the CCSARP treatment of apologies, see Olshtain (1989). Other studies of apologies include Holmes (1990), Bergman and Kasper (1993), Jaworski (1994), and Ogiermann (2009). The original CCSARP team collected their data by using discourse completion tasks (DCTs; see 9.3), which tended to produce longer pieces of speech than the single-utterance examples in the speech-act writings of Searle and others, but are apt to encourage briefer apologies (or other speech events) than, say, role plays (see 9.4, and Chapter 9 in general on these and other data collection methods).

(2) Modifiers as discussed here are not necessarily modifiers in the standard grammatical sense, although in the case of adverbs such as very, really, etc., they clearly are.

(3) On apology as a prototype category, see Jaworski (1994) and Deutschmann (2003).

(4) Kerbrat-Orecchioni (2006: 84) observes, with reference to French, that thanks are often upgraded (Merci beaucoup / mille fois / infiniment ‘Thank you very much / a thousand times / infinitely’) but never downgraded, except, significantly, when referring to thanks addressed by O to S: Tu pourrais me dire un petit merci ‘You could say a little thank you [to me].’ Similar observations could be made about English.

(5) But some polite speech acts, such as apologies, have a mixture of face threat and face enhancement. By acknowledging a fault, the apologizer admits to some loss of face to himself or herself. See 5.7.

(6) Blum-Kulka et al. (1989: 277) in the CCSARP coding scheme define alerters as “element[s]‌ whose function is to alert the Hearer’s attention to the ensuing speech act,” and they place Excuse me alongside Hey in this function. However, it is clear that the politeness function of Excuse me is not totally eclipsed by its attention-getting function. Eliciting the attention of someone in order to engage them in conversation is making a demand on them that often requires an apologetic stance, as shown in Excuse me or more elaborate apologies: Sorry to bother you, but.... By contrast, Hey (as in Hey, what’s the name of this street?) can seem rude and intrusive.

(7) In the spoken BNC, there are only twelve examples of Excuse me(, ) please, and no examples of Pardon me(, ) please. An example of the former is: Excuse me please, I’m trying to cook. Another pointer to the relative mildness of Excuse me is that, unlike I’m sorry, it cannot be intensified: there is no *Excuse me very much comparable to I’m very sorry.

(8) Compare Searle’s remark, in dealing with the related speech act of thanking, that “sincerity and essential rules overlap.”

(9) Bleaching in the sense of semantic generalization or weakening is characteristic of both grammaticalization and pragmaticalization as historical processes (see Hopper and Traugott 2003).

(10) Nevertheless, What? is by far the most common request for repetition in the conversational corpora BNCdemog and LCSAE. In the context-governed subcorpus of the BNC, containing more public and formal speech, Pardon? is commoner than What? reflecting its association with more polite usage. As requests for repetition, Pardon? is less common in AmE than in BrE, whereas Excuse me? is more common in AmE than in BrE (where, indeed, it is scarcely used).

(11) Sorry? is much less common as a repetition request in AmE, where in contrast Excuse me? has higher frequency.